Sensory evaluation and acceptability of quinoa–stevia cookies: Promoting sustainable food choices for nutritional and social development

Introduction

Consumer interest in nutritionally oriented bakery products

Public health guidance has moved decisively towards limiting free sugars; the World Health Organization recommends that free sugars supply less than 10% of total energy intake, with further benefit expected below 5% (WHO, 2015) [1] . Sweet bakery products are an obvious target for reformulation because they are eaten frequently and rely on sucrose for more than sweetness. The practical difficulty is that consumers rarely trade enjoyment for nutritional intent: a reformulated product that tastes noticeably different is unlikely to be purchased twice. Reviews of sugar-reduction strategies conclude that large reductions without sensory penalties are hard to achieve, particularly when high-intensity sweeteners introduce bitter side-tastes or altered temporal sweetness profiles (Hutchings et al., 2019) [2] . For this reason, sensory acceptance is not a secondary check in reformulation work but the criterion on which the whole exercise stands or falls.

Cookies as a vehicle for food product innovation

Cookies combine long shelf life, portion control and wide appeal across age groups, which makes them a convenient carrier for alternative flours and sweeteners. Their low-moisture, sugar-and fat-rich dough develops only a limited gluten network, so the formulation tolerates substantial changes to the flour fraction (Pareyt et al., 2008) [3] . Sucrose, however, is far from inert in this system. It competes with flour components for water, delays dough setting, governs spread and surface cracking, and influences the internal cell structure and therefore the crispness of the baked product (Pareyt et al., 2008; Pareyt et al., 2009) [3,4] . Replacing sucrose with a sweetener used at a small fraction of its mass removes bulk as well as sweetness, which is why sugar replacement in sweetened bakery goods typically requires attention to texture, volume and colour as well as taste (Struck et al., 2014) [5] .

Quinoa as a nutritionally valuable pseudocereal

Quinoa (Chenopodium quinoa Willd.) is an Andean pseudocereal whose grain is valued for protein of favourable amino acid balance, including a higher lysine content than common cereals, together with minerals, dietary fibre, unsaturated lipids and phenolic compounds (Vega-Gálvez et al., 2010; Navruz-Varli et al., 2016) [6,7] . Being gluten-free, it has attracted interest as an ingredient for improving the nutritional profile of bakery and gluten-free products (Alvarez-Jubete et al., 2010) [8] . Its incorporation into bread and cookies has been shown to raise lysine and other essential amino acids in the finished product (Stikic et al., 2012; Watanabe et al., 2014) [9,10] . Cultivation has also spread well beyond the Andes, including trials and production in India, and the crop's tolerance of marginal conditions has made it a candidate for agricultural diversification (Bhargava et al., 2006; Jacobsen, 2003; Bazile et al., 2016) [11,12,13] . Quinoa is not, however, sensorially neutral. Saponins concentrated in the seed pericarp contribute bitterness and astringency, and quinoaenriched products can suffer reduced acceptance unless the grain is washed, pearled or otherwise processed (Suárez-Estrella et al., 2018) [14] . In cookies, substituting 15% of wheat flour with quinoa flour lowered flavour, taste and overall liking relative to a wheat control, even though the product remained acceptable (Watanabe et al., 2014) [10] . Any strategy that improves the palatability of quinoa-containing cookies is therefore of direct practical relevance.

Stevia and sugar reduction

Stevia rebaudiana Bertoni produces steviol glycosides, principally stevioside and rebaudioside A, which are roughly 100-300 times sweeter than sucrose and contribute negligible energy at the concentrations used (Goyal et al., 2010; Carakostas et al., 2008) [15,16] . Purified rebaudioside A is heat-stable and has been extensively evaluated for use in foods and beverages (Carakostas et al., 2008) [16] . The sensory limitation of steviol glycosides is well documented: they activate the bitter taste receptors hTAS2R4 and hTAS2R14, giving a bitter or liquorice-like aftertaste, and some glycosides show a decline in sweetness at high concentration (Hellfritsch et al., 2012) [17] . Consequently, their performance depends strongly on dose and on the food matrix, and stevia should be regarded as a formulation tool whose acceptability must be demonstrated rather than assumed. In bakery systems, partial rather than complete sucrose replacement has generally been more successful. Rebaudioside A combined with fibres permitted partial sucrose replacement in muffins (Zahn et al., 2013) [18] , and in biscuits a 50:50 sucrose-stevioside ratio received higher sensory scores than either full sugar or complete replacement (Vatankhah et al., 2015) [19] . Sugar-reduced short-dough biscuits more broadly show changes in hardness, spread and liking that depend on the replacer used (Rodriguez-Garcia et al., 2022) [20] .

Importance of sensory evaluation

Hedonic testing asks assessors how much they like a product rather than how intense its attributes are, and the nine-point hedonic scale remains the most widely used instrument for this purpose (Peryam et al., 1957; Lawless et al., 2010) [21,22] . ISO 11136 provides general guidance on conducting hedonic tests with consumers in a controlled area (ISO, 2014) [23] . For cookies, liking of colour and appearance governs first impressions, flavour and taste capture the sweetener's contribution and any off-notes, and texture, crispiness and mouthfeel reflect the structural role of sucrose. Rating these attributes alongside overall acceptability allows the source of any change in liking to be located.

Research gap

Studies of quinoa cookies have focused chiefly on flour substitution (Watanabe et al., 2014) [10] , whereas studies of stevia in baked goods have generally used wheat-based formulations (Zahn et al., 2013; Vatankhah et al., 2015) [18,19] . Much less is known about how graded sucrose replacement behaves in a cookie that already contains a substantial proportion of quinoa flour, where the sweetener could interact with the grain's own flavour. In addition, hedonic data from such studies are frequently analysed with one-way ANOVA even when every panelist tastes every sample, an approach that ignores the repeated-measures structure of the design. The present work addresses both points.

Research objective

The objective of this study was to evaluate the effect of replacing 0-30% of sucrose with a stevia-based sweetener on the sensory acceptability of cookies containing 30% quinoa flour in the flour blend, to identify the formulation receiving the highest overall acceptability, and to describe how individual sensory attributes relate to overall liking.

Materials and Methods

Study design

The study formed the sugar-replacement stage of a sequential formulation programme in which the quinoa flour level had been fixed at 30% of the flour blend in an earlier stage. Seven formulations differing only in the sucrose:sweetener ratio were evaluated by the same 20 panelists, each of whom rated all seven samples for eight attributes. The design is therefore a balanced complete block, with panelists as blocks and formulations as the within-panelist factor.

Raw materials

The formulation comprised refined wheat flour (maida), quinoa flour, sucrose, a stevia-based sweetener, fat, baking powder and baking soda. Because the sweetener replaced sucrose on an equal-weight basis (Table 1), its composition determines the sweetness delivered and must be stated precisely.

Cookie preparation

The cookies were prepared according to the formulations presented in Table 1. The required ingredients for each formulation were weighed separately and combined using a conventional creaming and mixing procedure. The fat and sweetening ingredients were first mixed to obtain a uniform cream, followed by the gradual incorporation of the remaining dry ingredients. The mixture was blended until a homogeneous dough of suitable consistency was obtained. The prepared dough was subsequently sheeted to a uniform thickness and cut into circular pieces using a cookie cutter of consistent diameter. The shaped dough pieces were arranged on baking trays and baked in a preheated oven at 180 °C for 25 min, as specified in the formulation records. Baking was continued until the cookies developed the desired structural and surface characteristics. Following baking, the cookies were removed from the oven and allowed to cool to room temperature under ambient conditions. After cooling, the samples were packaged appropriately to minimize exposure to environmental moisture and other factors that could influence their sensory characteristics. The packaged cookies were maintained under the specified conditions until sensory evaluation.

Experimental formulations

The reference formulation (C 3 S 0 ) contained 70 g refined wheat flour, 30 g quinoa flour, 40 g sucrose, 50 g fat, 2 g baking powder and 2 g baking soda. In C 3 S 1 -C 3 S 6 , sucrose was reduced in 2 g steps and replaced by the same mass of stevia-based sweetener, corresponding to 5-30% replacement of sucrose by weight (Table 1). All other ingredients were held constant. C 3 S 0 is the full-sugar reference; the 30% quinoa level is common to all samples. Sucrose replaced (%) = sweetener mass ÷ 40 g × 100.

Sensory evaluation

Sensory evaluation was conducted using 20 panellists (P01-P20), each of whom evaluated all seven cookie formulations. The samples were assessed for colour, flavour, taste, texture, crispiness, mouthfeel, and overall acceptability using a conventional nine-point hedonic scale (1 = dislike extremely; 9 = like extremely) (Peryam et al., 1957) [21] . Samples were presented using random three-digit codes under comparable conditions, and panellists were provided drinking water for palate cleansing between samples. Individual scores were recorded for each attribute and formulation and subsequently used to calculate mean sensory scores and variability for statistical comparison of the developed quinoa-stevia cookies.

Sensory attributes

Eight attributes were rated: colour, appearance, flavour, taste, texture, crispiness, mouthfeel and overall acceptability. All eight present in the dataset were analysed.

Statistical analysis

The raw dataset was first audited for completeness, duplicate records, out-of-range values and consistency between the long-format and wide-format worksheets and the pre-computed summary sheets.

Descriptive statistics (mean, standard deviation, median) were calculated for each formulation × attribute combination (n = 20 per cell). Because every panelist rated every formulation, observations are not independent, and one-way ANOVA was therefore inappropriate. Hedonic scores are ordinal with a restricted range, so the Friedman test, a rank-based analysis for complete block designs, was used as the primary test for each attribute (Friedman, 1937) [24] . Kendall's coefficient of concordance, W = χ 2 / [n(k -1)], was reported as the effect size, indicating the degree of agreement among panelists in ranking the formulations (Kendall et al., 1939) [25] . Where the Friedman test was significant, all 21 pairwise comparisons were made with Wilcoxon signed-rank tests (Pratt treatment of zero differences, normal approximation), and p-values were adjusted within each attribute using the Holm procedure [26] . Significance groupings were summarised as compact letter displays, which were checked against the adjusted pairwise results. Matched-pairs rank-biserial correlation (r rb ) and the median and mean paired difference were reported as effect sizes for key comparisons. As a sensitivity analysis, one-way repeated-measures ANOVA with Greenhouse-Geisser correction was also fitted for each attribute. To test whether the response to replacement level was non-monotonic, overall acceptability was modelled with a linear mixed model containing linear and quadratic terms for sweetener level and a random intercept for panelist; linear and quadratic models were compared with a likelihood-ratio test. Associations between attributes were estimated with repeated-measures correlation (r rm ), which removes between-panelist differences in scale use and so avoids treating the 140 non-independent observations as independent (Bakdash et al., 2017) [27] . Principal component analysis was considered but not performed, because the seven formulation means across eight highly inter-correlated attributes provide too few observations for a stable or informative multivariate solution. Analyses were carried out in Python 3 (SciPy, pingouin, statsmodels); figures were generated with Matplotlib directly from the raw scores.

Significance was set at 𝛼 = 0.05.

Results

Data quality and descriptive statistics

The dataset contained 1,120 scores (20 panelists × 7 formulations × 8 attributes) with no missing values, duplicate panelist-sample-attribute records or scores outside the scale. Long-and wide-format records agreed completely, and the pre-computed means, standard deviations and Friedman statistics in the workbook matched independent recalculation. Scores ranged from 5 to 9 (5: 9 scores; 6: 157; 7: 445; 8: 368; 9: 141); no formulation received a score below 5, indicating that none was disliked by any panelist. Mean scores per panelist ranged from 6.68 to 8.41, showing some individual differences in scale use that the within-panelist analyses account for. Means and standard deviations are presented in Table 2 and Figure 1.

Colour and appearance

Colour liking increased from 6.65 in C 3 S 0 to 8.50 in C 3 S 4 and then declined to 7.40 in C 3 S 6 (Friedman χ 2 (6) = 83.90, p < 0.001, W = 0.70). C 3 S 4 was rated significantly higher than every other formulation for colour. Appearance followed the same pattern (6.45 to 8.40, then 7.35; χ 2 (6) = 84.13, p < 0.001, W = 0.70), although for appearance C 3 S 4 and C 3 S 3 did not differ significantly. Visual attributes showed the smallest decline between 20% and 30% replacement (-1.10 for colour, -1.05 for appearance).

Flavour

Flavour scores rose from 6.35 (C 3 S 0 ) to a maximum of 8.10 (C 3 S 4 ) and fell to 6.80 at 30% replacement (χ 2 (6) = 85.21, p < 0.001, W = 0.71). C 3 S 4 did not differ significantly from C 3 S 3 (7.85). Notably, flavour liking of C 3 S 6 was not significantly different from that of the full-sugar reference, making flavour the only attribute for which the 30% formulation lost its advantage over C 3 S 0 entirely.

Taste

Taste produced a large formulation effect (χ 2 (6) = 92.78, p < 0.001, W = 0.77). Mean scores increased from 6.35 to 8.45 between C 3 S 0 and C 3 S 4 , the largest rise of any attribute apart from overall acceptability and mouthfeel (+2.10), and then fell to 7.05 at 30% replacement. C 3 S 4 and C 3 S 3 (8.05) formed the top group; C 3 S 6 remained significantly above C 3 S 0 .

Texture, crispiness and mouthfeel

The textural attributes responded in the same inverted-U manner. Texture rose from 6.45 to 8.20 (χ 2 (6) = 80.75, W = 0.67), crispiness from 6.50 to 8.30 (χ 2 (6) = 84.43, W = 0.70) and mouthfeel from 6.55 to 8.60 (χ 2 (6) = 95.43, W = 0.80), all p < 0.001. For each of the three, C 3 S 4 was significantly higher than all other formulations. Mouthfeel showed the steepest fall beyond the peak, from 8.60 at 20% to 7.00 at 30% replacement.

Overall acceptability

Overall acceptability showed the strongest formulation effect and the highest panel agreement of all attributes (χ 2 (6) = 100.58, p < 0.001, W = 0.84). C 3 S 4 received the highest mean overall acceptability (8.65 ± 0.49; median 9), followed by C 3 S 3 (8.35 ± 0.59), C 3 S 5 (7.80 ± 0.52) and C 3 S 2 (7.65 ± 0.75); C 3 S 1 (7.10 ± 0.72) and C 3 S 6 (7.05 ± 0.51) formed a lower group, and the full-sugar reference C 3 S 0 scored lowest (6.40 ± 0.60) (Figure 2A). Every one of the 20 panelists gave C 3 S 4 a higher overall acceptability score than C 3 S 0 , and C 3 S 4 was the highest-rated or joint-highest sample for all 20 panelists. Thirteen panelists (65%) awarded C 3 S 4 the maximum score of 9, whereas 13 panelists (65%) scored C 3 S 0 at 6 (Figure 2B). C 3 S 4 was significantly preferred to every other formulation (Table 4). The difference from C 3 S 3 must, however, be interpreted with care: the mean paired difference was only 0.30 points, 14 of the 20 panelists gave the two samples identical scores, and the Holm-adjusted pvalue (0.043) lay close to the significance threshold. C 3 S 3 and C 3 S 4 are therefore best regarded as the two most accepted formulations, with C 3 S 4 having the highest mean under the conditions of this study.

Statistical comparisons and robustness

Friedman results for all attributes are summarised in Table 3. Panel agreement was high throughout (W = 0.67-0.84), meaning that panelists ranked the formulations in broadly the same order. Repeatedmeasures ANOVA led to identical conclusions for every attribute (all Greenhouse-Geisser-corrected p 0.001; generalised η 2 = 0.42-0.60), confirming that the findings do not depend on the choice between rank-based and parametric analysis.

Relationships among sensory attributes

All attributes were positively and strongly associated within panelists (r rm = 0.58-0.77, df = 119, all p 0.001; Figure 3). Overall acceptability was most closely associated with flavour (r rm = 0.77), mouthfeel (0.75), taste (0.74) and appearance (0.73), and less closely with tex-ture (0.71), crispiness (0.69) and colour (0.66). The weakest association was between colour and texture (0.58). The narrow spread of coefficients indicates that panelists who liked a sample more tended to rate all of its attributes higher, so the correlations should not be read as evidence that any single attribute causes overall liking.

Multivariate analysis

Principal component analysis and preference mapping were not undertaken. With seven formulations and eight attributes that are strongly collinear (Figure 3), a PCA of mean scores would be dominated by a single overall-liking dimension already evident in Table 2 and Figure 1, and preference mapping requires descriptive or instrumental data that were not collected.

Discussion

Influence of quinoa on sensory characteristics

Because all seven formulations contained 30% quinoa flour, the present data cannot isolate the effect of quinoa on liking. What they do show is the starting point: the full-sugar 30%-quinoa reference was rated between like slightly and like moderately (6.40), a level broadly consistent with earlier reports that quinoa substitution in cookies reduces flavour and taste liking while leaving the product acceptable (Watanabe et al., 2014) [10] . Bitter and astringent saponins are the most commonly cited cause of this penalty (Suárez-Estrella et al., 2018) [14] . That every stevia-containing formulation up to 25% replacement exceeded the reference in all attributes suggests that the sugar-sweetener balance can partly offset the sensory limitations of a quinoa-rich base, although this interpretation remains to be tested directly against a quinoa-free control.

Influence of stevia on sensory acceptance

The most notable result is that moderate replacement of sucrose increased rather than decreased liking. This departs from the common expectation that sugar reduction carries a sensory cost (Hutchings et al., 2019) [2] , and it needs explanation. The data do not identify the mechanism, but several are plausible. First, the stevia-based sweetener was added on an equal-weight basis; because steviol glycosides are far sweeter than sucrose (Goyal et al., 2010; Carakostas et al., 2008) [15,16] , replacement may have increased perceived sweetness or improved the sweetness-bitterness balance of the quinoa base, depending on the sweetener's exact composition. Second, reducing sucrose alters dough setting, spread and cell structure (Pareyt et al., 2008; Pareyt et al., 2009) [3,4] , and moderate changes of this kind could have produced a texture or appearance that panelists preferred. Third, the strong inter-attribute correlations point to a general liking effect, in which a sample favoured for its taste is also rated more favourably for colour and texture (Lawless et al., 2010) [22] . The decline beyond 20% replacement is consistent with known properties of steviol glycosides. As concentration rises, their bitter and liquorice-like notes, mediated by hTAS2R4 and hTAS2R14, become more prominent, and sweetness may plateau or decline (Hellfritsch et al., 2012) [17] . In this study the sharpest losses beyond the peak were in mouthfeel, overall acceptability, taste and flavour, while colour and appearance changed least, a pattern that fits an oral rather than a visual cause. Loss of sucrose bulk at higher replacement may also have contributed to the textural decline (Struck et al., 2014) [5] . Instrumental texture, colour and sweetness-intensity measurements would be needed to separate these explanations.

Relationship between sensory attributes and overall acceptability

Flavour, mouthfeel and taste were the attributes most closely associated with overall acceptability, suggesting that oral perception carried more weight in liking than visual attributes did. The associations are correlational and, given how uniformly high they were, partly reflect panelists' general liking of each sample. They identify where formulation effort is most likely to pay off, namely controlling flavour and aftertaste at higher replacement levels, but they do not establish causal drivers.

Comparison with previous research

The inverted-U response agrees with work on stevioside in biscuits, where an intermediate 50:50 sucrose-stevioside ratio outperformed both full sugar and complete replacement (Vatankhah et al., 2015) [19]

, and with reports that rebaudioside A performs best as a partial replacer in baked goods (Zahn et al., 2013) [18] . It agrees partially with reviews concluding that large sugar reductions are difficult to achieve without sensory penalties (Hutchings et al., 2019; Struck et al., 2014) [2, 5]: the penalty appeared here only above 20%, and the replacement range tested was modest compared with complete-replacement studies. The finding that partial replacement actually raised liking above the full-sugar reference differs from much of the literature, in which the best-case outcome is usually parity with the control (Rodriguez-Garcia et al., 2022) [20] . Possible reasons include the quinoa-rich base, the equal-weight addition of a high-potency sweetener, and panel characteristics that are not yet documented (Section 5). The present results for quinoa cookies also differ from those of (Watanabe et al.) [10] in scale and design, as that study varied quinoa rather than sweetener and used a seven-point bipolar scale, so direct numerical comparison is not appropriate.

Product development implications

For product development, the data support replacing about 15-20% of sucrose with the stevia-based sweetener in this quinoa cookie base, with C 3 S 4 (20%) giving the highest mean acceptance and C 3 S 3 (15%) performing almost as well. Replacement of 25% or more reduced liking across all attributes and should not be adopted without reformulation, for example by adding bulking agents or flavour-masking strategies (Hutchings et al., 2019; Struck et al., 2014) [2,5] . Because liking was consistently high and panel agreement strong, C 3 S 4 is a reasonable base for the next stage of formulation work; its sugar content should nonetheless be verified analytically before any reduced-sugar claim is made.

Sustainability and social development perspective

This study measured sensory liking only. It did not measure nutritional composition, health outcomes, environmental impact, cost, or consumer behaviour, and its results cannot establish benefits in any of these areas. Externally, quinoa has been promoted as a crop for agricultural diversification, and its cultivation has expanded into many new regions, including India (Bhargava et al., 2006; Bazile et al., 2016) [11,13] ; lowering free-sugar intake is also a recognised publichealth goal (WHO, 2015) [1] . Acceptable reduced-sugar quinoa products are a necessary precondition for such benefits, because foods that are not eaten deliver nothing. Whether the formulations tested here contribute to sustainable food choices or to nutritional and social development is a question for future life-cycle, nutritional and consumer studies, not a finding of the present work.

Limitations

The study has several limitations that should be considered when interpreting the findings. The sensory panel was relatively small (n = 20), which may limit the generalisability of the results to broader consumer populations. The quinoa incorporation level was fixed at 30%, preventing assessment of the independent effect of different quinoa levels on sensory acceptance. In addition, the study focused exclusively on hedonic sensory responses and did not include descriptive sensory profiling or instrumental measurements of texture, colour, or sweetness intensity; therefore, explanations regarding the sensory mechanisms underlying the observed preferences should be interpreted cautiously. Nutritional composition, sugar reduction, shelf life, purchase intention, and environmental impacts were also outside the scope of the present investigation. Future studies should therefore employ larger and more diverse consumer panels and integrate sensory, instrumental, nutritional, and consumer-behavioural measurements to provide a more comprehensive evaluation of quinoa-stevia cookies.

Conclusion

Graded replacement of sucrose with a stevia-based sweetener had a strong and consistent effect on the sensory acceptability of cookies containing 30% quinoa flour. Liking of all eight attributes increased up to 20% replacement and declined at 25-30%, with high agreement among panelists. The 20% formulation (C 3 S 4 ) received the highest mean score for every attribute and the highest overall acceptability (8.65 ± 0.49), significantly exceeding all other formulations, although its advantage over the 15% formulation was small. Flavour, mouthfeel and taste were the attributes most closely associated with overall liking. Under the conditions of this study, partial sucrose replacement in the range of 15-20% is therefore the most promising option for further development of this quinoa cookie. Nutritional, health, environmental and social benefits were not measured and should not be inferred from these results.

References

1. World Health Organization. Guideline: sugars intake for adults and children. Geneva: WHO; 2015.
2. Hutchings SC, Low JYQ, Keast RSJ. Sugar reduction without compromising sensory perception. An impossible dream? Crit Rev Food Sci Nutr. 2019;59(14):2287-2307.
3. Pareyt B, Delcour JA. The role of wheat flour constituents, sugar, and fat in low moisture cereal based products: a review on sugarsnap cookies. Crit Rev Food Sci Nutr. 2008;48(9):824-839.
4. Pareyt B, Talhaoui F, Kerckhofs G, Brijs K, Goesaert H, Wevers M, Delcour JA. The role of sugar and fat in sugar-snap cookies: structural and textural properties. J Food Eng. 2009;90(3):400408.
5. Struck S, Jaros D, Brennan CS, Rohm H. Sugar replacement in sweetened bakery goods. Int J Food Sci Technol. 2014;49(9):1963-1976.
6. Vega-Gálvez A, Miranda M, Vergara J, Uribe E, Puente L, Martínez EA. Nutrition facts and functional potential of quinoa (Chenopodium quinoa Willd.), an ancient Andean grain: a review. J Sci Food Agric. 2010;90(15):2541-2547.
7. Navruz-Varli S, Sanlier N. Nutritional and health benefits of quinoa (Chenopodium quinoa Willd.). J Cereal Sci. 2016;69:371376.
8. Alvarez-Jubete L, Arendt EK, Gallagher E. Nutritive value of pseudocereals and their increasing use as functional gluten-free ingredients. Trends Food Sci Technol. 2010;21(2):106-113.
9. Stikic R, Glamoclija D, Demin M, Vucelic-Radovic B, Jovanovic Z, Milojkovic-Opsenica D, Jacobsen SE, Milovanovic M. Agronomical and nutritional evaluation of quinoa seeds (Chenopodium quinoa Willd.) as an ingredient in bread formulations. J Cereal Sci. 2012;55(2):132-138.
10. Watanabe K, Kawanishi-Asaoka M, Myojin C, Awata S, Ofusa K, Kodama K. Amino acid composition, oxidative stability, and consumer acceptance of cookies made with quinoa flour. Food Sci Technol Res. 2014;20(3):687-691.
11. Bhargava A, Shukla S, Ohri D. Chenopodium quinoa-an Indian perspective. Ind Crops Prod. 2006;23(1):73-87.
12. Jacobsen SE. The worldwide potential for quinoa (Chenopodium quinoa Willd.). Food Rev Int. 2003;19(1-2):167-177.
13. Bazile D, Jacobsen SE, Verniau A. The global expansion of quinoa: trends and limits. Front Plant Sci. 2016;7:622.
14. Suárez-Estrella D, Torri L, Pagani MA, Marti A. Quinoa bitterness: causes and solutions for improving product acceptability. J Sci Food Agric. 2018;98(11):4033-4041.
15. Goyal SK, Samsher, Goyal RK. Stevia (Stevia rebaudiana) a biosweetener: a review. Int J Food Sci Nutr. 2010;61(1):1-10.
16. Carakostas MC, Curry LL, Boileau AC, Brusick DJ. Overview: the history, technical function and safety of rebaudioside A, a naturally occurring steviol glycoside, for use in food and beverages. Food Chem Toxicol. 2008;46(7 Suppl):S1-S10.
17. Hellfritsch C, Brockhoff A, Stähler F, Meyerhof W, Hofmann T. Human psychometric and taste receptor responses to steviol glycosides. J Agric Food Chem. 2012;60(27):6782-6793.
18. Zahn S, Forker A, Krügel L, Rohm H. Combined use of rebaudioside A and fibres for partial sucrose replacement in muffins. LWT -Food Sci Technol. 2013;50(2):695-701.
19. Vatankhah M, Garavand F, Elhamirad A, Yaghbani M. Influence of sugar replacement by stevioside on physicochemical and sensory properties of biscuit. Qual Assur Saf Crops Foods. 2015;7(3):393-400.
20. Rodriguez-Garcia J, Ding R, Nguyen THT, Grasso S, Chatzifragkou A, Methven L. Soluble fibres as sucrose replacers: effects on physical and sensory properties of sugar-reduced shortdough biscuits. LWT. 2022;167:113837.
21. Peryam DR, Pilgrim FJ. Hedonic scale method of measuring food preferences. Food Technol. 1957;11(Suppl):9-14.
22. Lawless HT, Heymann H. Sensory evaluation of food: principles and practices. 2nd ed. New York: Springer; 2010.
23. International Organization for Standardization. ISO 11136:2014 Sensory analysis -Methodology -General guidance for conducting hedonic tests with consumers in a controlled area (incl. Amd 1:2020). Geneva: ISO; 2014.
24. Friedman M. The use of ranks to avoid the assumption of normality implicit in the analysis of variance. J Am Stat Assoc. 1937;32(200):675-701.
25. Kendall MG, Babington Smith B. The problem of m rankings. Ann Math Stat. 1939;10(3):275-287.
26. Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat. 1979;6(2):65-70.
27. Bakdash JZ, Marusich LR. Repeated measures correlation. Front Psychol. 2017;8:456.
28. How to cite this article: Manish, Kumar S, Namrata, Kumar R. Sensory evaluation and acceptability of quinoa-stevia cookies: Promoting sustainable food choices for nutritional and social development. International Journal of Agriculture Extension and Social Development. 2026;9(10):89-96. doi: https://www.doi.org/ 10.33545/26180723.2026.v9.i10b.4363